Jake VanderPlas is an astronomer and open-source leader, serving as Director of Open Software at the University of Washington’s eScience Institute. He writes and builds widely used Python tools for data science, machine learning, and scientific computing.
66 articles from this blog
A response to a critique of the author's introductory series on Frequentist vs. Bayesian statistics, focusing on audience and the role of decision theory.
A reflection on the challenges of data science in academia, discussing the 'brain drain' of data skills and the need for systemic change.
A practical guide to implementing Bayesian analysis in Python using MCMC packages like emcee, PyMC, and PyStan, with a line-fitting example.
Explores the critical difference between frequentist confidence intervals and Bayesian credible regions, arguing why frequentism often fails scientific inquiry.
A data scientist analyzes Seattle's bicycle counter data using Python to determine if cycling is truly increasing or just affected by good weather.
Explores differences between frequentist and Bayesian statistics, focusing on how they handle nuisance parameters in data analysis.
An in-depth technical explanation of why Python is slower than compiled languages, focusing on dynamic typing, interpretation, and object overhead.
Explains the Python Buffer Protocol (PEP 3118), showing how objects like arrays and NumPy ndarrays share data efficiently without copying.
A practical introduction to the philosophical and practical differences between frequentist and Bayesian statistics, with Python examples.
Introducing mpld3's plugin framework for creating interactive matplotlib visualizations in the browser using D3.js.
Introducing mpld3, a tool that creates interactive D3.js visualizations from matplotlib plots for use in IPython notebooks.
Introducing ipywidgets, a library for creating static, interactive widgets in IPython/Jupyter notebooks that work without a live kernel.
A comparison of four Python implementations for Kernel Density Estimation (KDE), analyzing their features, interfaces, and performance.
Explores how the demand for big data skills in industry is draining talent from academic science, threatening research.
An exploration of the Cooley-Tukey Fast Fourier Transform (FFT) algorithm, its underlying principles, and a Python implementation from scratch.
A tutorial on implementing Conway's Game of Life in Python using NumPy and SciPy, with visualization via matplotlib animations.
Matplotlib 1.3 introduces a built-in plt.xkcd() function to easily create plots in the style of the popular XKCD webcomic.
A performance comparison revisiting Numba vs. Cython for optimizing a pairwise distance array computation, using updated libraries and benchmarks.
Explains how to pass data between Javascript and Python in IPython/Jupyter notebooks using kernel.execute, though notes the method is now obsolete.
A blog post demonstrating how to create a puzzling shape animation using Python's matplotlib and IPython notebook.
